Platform Trials Are Neurology’s Best Bet — So Why Is the FDA Still Treating Them Like Experiments?

The Lancet’s neurology commentary landed this week with the kind of quiet authority that tends to move slowly through regulatory circles and fast through R&D strategy meetings: multi-arm multi-stage (MAMS) platform trials represent the most credible path forward for neurodegenerative diseases that have defeated conventional Phase 3 development for decades. The argument is not theoretical. Alzheimer’s disease, Parkinson’s disease, and motor neuron disease have collectively consumed billions in late-stage investment and returned almost nothing — and the trial architecture has been as much a culprit as the biology.

Consider what happened to Biogen and Eisai in March 2019. They terminated both Phase 3 trials of aducanumab — EMERGE and ENGAGE — after a futility analysis concluded the drug would not meet its primary endpoint. Hundreds of millions of dollars and years of patient time were consumed before the signal collapsed. A platform design with pre-specified interim stopping rules and shared control arms would have surfaced that futility signal earlier, with fewer patients exposed and faster resource reallocation to the next candidate.

That pattern — late-stage failure after inadequate early signal detection — is the central operational failure MAMS trials are designed to correct. The question for clinical operations leaders in 2026 is no longer whether the design is scientifically valid. The question is whether the regulatory and site infrastructure can actually support it at scale.

The Old Assumption Is No Longer Affordable

The legacy model assumed that neurodegenerative trials would proceed sequentially: one candidate, one Phase 2, one Phase 3, one regulatory submission. That model made sense when the mechanistic landscape was sparse. Sponsors picked their best hypothesis and ran with it. But the 2025 Alzheimer’s disease drug development pipeline now encompasses 182 clinical trials evaluating 138 candidate treatments — a number that has grown from 164 trials in 2024. Running those candidates through sequential, independent Phase 3 programs is arithmetically impossible at any responsible budget, and scientifically wasteful when shared biomarker infrastructure and control arms could generate comparative data simultaneously.

The new assumption — embedded in the Lancet commentary and increasingly visible in how sophisticated sponsors are structuring their pipelines — treats the platform as the unit of investment, not the individual compound. The Alzheimer’s Tau Platform (ATP) trial is the clearest current example: a master protocol designed to evaluate tau-directed therapies, alone or in combination with anti-amyloid agents, across multiple arms with adaptive interim analyses. That architecture allows a failing arm to be dropped without disrupting the platform, and a promising signal to trigger seamless expansion into a confirmatory stage without a new trial start.

The efficiency case is compelling in numbers. According to data from Applied Clinical Trials, a global top-20 drug development company reported saving over $70 million annually through adaptive dose-finding in Phase II combined with sample size adjustments and futility stopping in Phase III. That figure does not account for the downstream value of faster competitive intelligence and earlier go/no-go decisions — both of which have strategic value in a disease space where first-mover advantage in mechanism is real.

Where the Infrastructure Breaks

The MAMS design works on paper. In practice, three operational fault lines emerge the moment a sponsor tries to execute one in a neurological indication.

The first is the master protocol governance problem. A MAMS platform requires a central governing committee with pre-specified authority over arm additions, interim analyses, and stopping decisions. Most sites — and most IRBs — are not structured to process protocol amendments at platform speed. A published MAMS design for progressive multiple sclerosis, from Li, Leurent, Barkhof et al. in Neurology (May 2022), demonstrated that a platform targeting a hazard ratio of 0.75 for the primary final outcome could materially reduce required sample size versus sequential trials — but the design depended on continuous biomarker assessment and coordinated interim analyses that standard site monitoring workflows do not accommodate.

The second fault line is biomarker standardization. Neurodegenerative platform trials require shared intermediate endpoints — amyloid PET, tau CSF, neurofilament light chain — that are both sensitive enough to detect mechanistic signal at interim and accepted by regulators as action-worthy. The FDA’s CDER, in collaboration with the Critical Path Institute, has been developing Model-Informed Drug Development (MIDD) approaches for neurological conditions — but the gap between MIDD methodology and operational biomarker qualification standards remains wide enough that a sponsor cannot yet submit a MAMS master protocol and assume FDA alignment on what constitutes a valid interim decision trigger.

The third fault line is the most uncomfortable one: the FDA has not issued dedicated guidance on MAMS platform trial design for neurodegenerative diseases. Sponsors operating under the agency’s 2019 Adaptive Designs for Clinical Trials of Drugs and Biologics guidance can extract applicable principles, but that document was written with single-disease, single-compound adaptive trials in mind. A multi-arm platform with arm-level stopping rules, shared controls, and seamless Phase 2/3 transitions requires a regulatory conversation that is still happening informally — in Type B meetings, in pre-IND briefing documents — rather than in published, citable guidance that sites and IRBs can use to standardize their processes.

The Directive for Sponsors Moving Now

If you are structuring a MAMS platform for a neurodegenerative indication, do not wait for formal FDA guidance to appear before engaging the agency. Request a Type B meeting with CDER’s Division of Neurology I or II early in master protocol development — specifically to align on three items: the statistical framework for arm-level interim decisions, the biomarker qualification threshold that would trigger a seamless Phase 3 expansion, and the acceptable range of protocol amendments that can be processed without triggering a full IND amendment cycle. The absence of published guidance makes that early alignment non-optional, not optional. Platforms that have tried to finalize master protocol architecture and then seek FDA endorsement have paid for that sequence in clock cycles they could not recover.

Site selection strategy also changes under a platform model. A MAMS trial in Alzheimer’s or Parkinson’s disease demands sites with continuous biomarker infrastructure — amyloid PET capacity, certified CSF collection, NfL assay access — and governance structures capable of processing arm-level amendments without pausing enrollment across the entire platform. That is a materially different site feasibility filter than a standard Phase 3 site qualification checklist, and sponsors who apply the old filter will build platforms on site networks that cannot sustain the operational rhythm the design requires.

The 2025 Alzheimer’s pipeline carries 138 candidate treatments across 182 trials. Most of those trials are running sequentially, independently, and expensively. The platforms that establish regulatory alignment and site infrastructure now will not just complete faster — they will generate comparative mechanistic data that reshapes the entire field’s understanding of what actually works in neurodegenerative biology. The sponsors still building sequential Phase 3 programs in 2027 will be reading those results as competitors, not co-authors.

References

1. The Lancet — “Multi-arm multi-stage platform trials for neurological disease: accelerating progress”
2. PMC / Alzheimer’s Research — “2025 Alzheimer’s disease drug development pipeline: 182 clinical trials, 138 candidate treatments”
3. Applied Clinical Trials Online — “Adoption of Adaptive Trial Designs Poised to Accelerate” (top-20 pharma company: $70M annual savings)
4. Neurology (PubMed) — Li V, Leurent B, Barkhof F, et al., “Designing Multi-arm Multistage Adaptive Trials for Neuroprotection in Progressive Multiple Sclerosis,” May 3, 2022
5. FDA CDER — “Opportunities for Innovation in Drug Development for Neurological Conditions” (MIDD collaboration with Critical Path Institute)
6. Alzforum — “Biogen/Eisai Halt Phase 3 Aducanumab Trials,” March 21, 2019